Poker Hand Rankings Trainer

Drill hand rankings against a real evaluator, checked against all 2,598,960 five-card hands. Which hand wins, what a hand is, and how rare each one is.

0
Right
0
Asked
0
Streak

Why the order is what it is

HandFive-card handsChanceAbout one in
Straight flush 40 0.0015% 64,974
Four of a kind 624 0.02% 4,165
Full house 3,744 0.14% 694
Flush 5,108 0.20% 509
Straight 10,200 0.39% 255
Three of a kind 54,912 2.11% 47
Two pair 123,552 4.75% 21
Pair 1,098,240 42.26% 2
High card 1,302,540 50.12% 2

Out of 2,598,960 possible five-card hands. The ranking order is not a convention someone chose — it is exactly the order of these counts.

How this knows the answer

  • The evaluator is checked against arithmetic rather than examples. All 2,598,960 five-card hands are classified and the totals compared with counts derived combinatorially — 40 straight flushes, 624 fours of a kind, 5,108 flushes, and so on. If the nine counts are right and they sum correctly, nothing is being misclassified.
  • That matters because a hand evaluator is easy to write and easy to get subtly wrong. A wheel ranked by its ace, a flush placed above a full house, a kicker compared in the wrong order — none of those show up on the handful of hands anyone checks by hand.
  • Five, four, three, two, ace is a straight and it is the LOWEST one, so it is ranked by its five rather than its ace. Getting that backwards turns the worst straight in the deck into the best one.
  • The order of the categories is not a convention someone chose — it follows the counts. A full house beats a flush because there are 3,744 full houses and 5,108 flushes, so the full house is rarer. Every rank in the list is in the order of how hard it is to make.
  • Ties are real and are reported as such. Two players can hold exactly the same hand when the board plays, and an evaluator that always picks a single winner is wrong about a great many hands.

How to use

  1. Pick a drill: name the hand, or spot the winner.
  2. Answer, then read what the hand actually was.
  3. Check the table for how rare each hand really is.

Frequently asked questions

How do you know the answers are right?

The evaluator is checked by exhaustion rather than by examples. All 2,598,960 five-card hands are classified and the nine totals compared against counts derived from binomials — 40 straight flushes, 624 fours of a kind, 5,108 flushes and so on. If every count matches and they sum correctly, nothing is being misclassified anywhere.

Does a flush beat a full house?

No, the other way round. There are 5,108 flushes and only 3,744 full houses, so the full house is rarer and ranks higher. That is the whole logic of the ranking order — it is not a convention someone chose, it is exactly the order of how hard each hand is to make.

Is ace-two-three-four-five a straight?

Yes, and it is the LOWEST straight in the deck — ranked by its five rather than its ace. Getting that backwards turns the worst straight into the best one, which is one of the classic ways a hand evaluator goes wrong. Nothing wraps past the ace, so king-ace-two-three-four is not a straight at all.

What happens when two hands are the same?

They tie, and the drill says so. Suits never break a tie in poker, so two players can genuinely hold the same hand — and a trainer that always names a single winner would be wrong about a great many hands. "They tie" is one of the answers.

What is the difference between the five and seven card drills?

Seven cards is how hold’em works: two in your hand and five on the board, and you play the best five of the seven. It is noticeably harder, because the hand you have is hidden among two extra cards that do not belong to it.

How rare is a royal flush really?

Four hands in 2,598,960, which is about one in 650,000 five-card deals. The table shows the same figure for every category, computed from the same binomials the evaluator is tested against, so the numbers on the page and the answers in the drill cannot drift apart.

🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.